imates 
surface 
, and 
ages. 
entific 
mes of 
of the 
f it is 
| of the 
ns and 
data in 
nn and 
oratory 
ogy in 
end to 
oratical 
ine. In: 
jl. 602, 
er 1982. 
' direct 
Diego, 
on and 
Sports Dynamics of Carl Lowis through 100 m Race using Video Imagery 
Hirofumi CHIKATSU” , Masako TURUOKA** , Shunji MURAI** 
* Tokyo Denki University 
Faculty of Science and engineering 
Hatoyama Hikigun Saitama 355-03 Japan 
Tel +81-492-96-2911 
Fax +81-492-96-6501 
Institute of Industrial Science 
** 
University of Tokyo 
Abstract 
Dynamic tracking of moving object became possible by using video imagery. Especially in 
the field of sports dynamic analysis, video imagery gives useful information because of 
video camera takes 30 images per second. 
In this study, video images of world champion 
ship held in Tokyo, Japan in August, 1991 were analyzed from view of sports dynamics. 
Gold medalist Carl Lowis was selected for time-series digital photogrammetric analysis. 
Keyword: Video imagery, Image Processing, 
1. Introduction 
Generally, video imagery provided by 
TV program includes many noises and there 
are no control points for orientation 
except a special case. Furthermore, it's 
must be considered that object and camera 
angle are always moving in the case of 
tracking of moving object. 
Therefore, to utilize video imagery for 
dynamic tracking of moving object by the 
Digital Photogrammetry, there are two big 
subjects. One is an image processing and 
the other is a photogrammetric problem. 
On the other hand to analysis a dynamics 
of human movement, it's necessary to 
recognize a feature points of man such as 
knees, elbows, feet and etc. by an image 
processing technique. 
The video images of World Champion Ship 
held in Tokyo 1991, have not any control 
point and these images were taken with a 
narrow angle lens. There are some 
orientation methods for CCD camera 
without control pointi)2) or tracking 
System for moving object with CCD camera 
mounted on a motorized theodolite9), but 
in this study these method or system 
can't be utilized because of the video 
images was coped from TV. Orientation 
  
Sports Dynamics 
method for these images has to be 
developed. 
This paper discusses about two 
complicated subjects through the video 
images of 100 m race of Carl Lowis and 
also shows the results of sports dynamics 
of Carl Lowis. 
2. Image Processing 
To analysis the dynamics of human 
movement by using video imagery, the 
first annoying subject is playback of 
clear imagery from noisy imagery. For 
example, Fig.2 shows an original imagery 
that Carl Lowis is running. Many white 
noise can be detected around him, these 
white noise gives big influence to the 
next feature extraction step. Next is 
recognition of feature points of man such 
as knees, elbows, etc. and automated 
positioning of the feature points. 
Fig.l shows image processing procedure by 
Personal Computer which used in this 
study. 
The detail procedures of each step in 
Fig.1 are as follows,